317 - From Portal Enrollment to Active Use: How Electronic Health Literacy Influences Caregiver Use of Patient Portals

Publication Number: 1301.317

Glenn Rosenbluth, University of California, San Francisco, School of Medicine, San Francisco, CA, United States; Danea Horn, University of California, San Francisco, San Francisco, CA, United States; Aris Oates, University of California, San Francisco, School of Medicine, San Francisco, CA, United States; Leslie Magana, UCSF Benioff Children's Hospital Oakland, Oakland, CA, United States; Alicia Fernandez, University of California, San Francisco, San Francisco, CA, United States; Naomi Bardach, University of California, San Francisco, School of Medicine, San Francisco, CA, United States

Background: Use of patient portals in electronic health records has been linked to improved clinical outcomes. Many quality improvement efforts focus on portal enrollment, but little is known about contributors to portal usage after enrollment. The electronic health literacy (eHL) framework describes cognitive, technological, communication, and social/environmental domains which may contribute to portal use. eHL domains can be assessed using the electronic health literacy questionnaire (eHLQ).

Objective: To assess associations between caregiver eHLQ scores and portal activation (enrolled before or during admission) and login-days during pediatric hospitalizations.

Design/Methods: English- and Spanish-speaking caregivers of inpatients < 12yo who participated in the Family Input for Quality and Safety (FIQS2) trial (8/2021-2/2024) completed demographic surveys and items (1-4 Likert scale) of 3 relevant eHLQ domains: “Ability to actively engage with digital services” (Ability); “Feel safe and in control with digital services” (Safe/In Control); and “Motivated to engage with digital services” (Motivation). We retrospectively reviewed proxy (i.e., caregiver) portal activation status (yes/no) during hospitalization and portal login-days (#inpatient days with at least one login). We examined associations between eHLQ domain scores and portal metrics using univariate and multivariate Poisson regression. Multivariate models controlled for length of stay, race, ethnicity, language (REaL), gender, and education.

Results: 1038 caregivers completed the eHLQ. In the simple adjusted model only containing 3 eHLQ domains, each domain was associated with both proxy activation and login-days (Tables 2 & 3). In the fully adjusted model, Motivation was the only eHLQ domain that was statistically significantly associated with proxy activation (OR 1.50, 95% CI [1.10,2.07], p=0.01) or proxy login-days (IRR 1.37, 95% CI [1.07,1.76], p=0.014). Across all models, Latinx Spanish-speakers were less likely to have proxy active (OR 0.30, 95% CI [0.16,0.57], p< 0.001) as were Black/African Americans (OR 0.47,95% CI [0.22,0.99], p< 0.05). Black/African Americans had fewer login-days (IRR 0.183, 95% CI [0.05,0.67], p=0.01).

Conclusion(s): Our simple adjusted model with eHLQ domains suggests that each domain of Ability, Safe/In Control, and Motivation are potential areas for interventions to increase engagement. The fully adjusted model suggests that Motivation is a potential area for universal intervention, and that future studies could explore specific interventions in Ability and Safe/In Control for groups with variation in the model (e.g., REaL).

Table 1: Participant Characteristics with Proxy Activation and Login Days
Table 1 MyChart eHLQ.pdfNote: Values are n (%) unless otherwise specified. eHLQ domains scores are an average of the 5 items (1-4 Likert scale) within each domain.

Table 2: Association between EHLQ Scores & MyChart Proxy Activation
Table 2 MyChart eHLQ.pdfNote: Logistic regression models reporting odds ratios (95% CI). All models control for log(patient-days), since the number of days is associated with greater opportunity to enroll in the portal. The fully adjusted model includes controls for race/ethnicity/language (REaL), gender, and education level. Reference group for REaL is “Declined to state” and for Education is Less than High School. * p<0.05, ** p<0.01, *** p<0.001.

Table 3: Association between EHLQ Scores & Count of MyChart Proxy Login-Days During Admission
Table 3 MyChart eHLQ.pdfNote: Poisson regression models reporting incidence rate ratios (95% CI). All models control for log(patient-days), since the number of days is associated with greater opportunity to log-in. The fully adjusted model includes controls for race/ethnicity/language (REaL), sex, and education level. Reference group for REaL is “Declined to state” and for Education is Less than High School. * p<0.05, ** p<0.01, *** p<0.001.